In a general copying apparatus, such as a photocopier, the original image placed on an original platen is irradiated with light from a light source, and the reflection light from the original image is converted into scan-image-information by an optical sensor, and the scan-image-information is subjected to various image compensations before outputted. However, it is still difficult to obtain scan-image-information exactly identical to the original image by the copying apparatus, because the reflection of the original image is not even at all times. Also, resolution for scanning the original image is limited; for example, small characters are often crushed in the scan-image-information.
Most of the original image placed on the original platen to be scanned is image information created in a computer and outputted by a printer or the like. Therefore, if the original image information exists, image quality is better for an output of the original information than the output of scan-image-information obtained through optical scanning of the original image by a copying apparatus.
As disclosed in Document 1 through 3, a particular image forming device is suggested in this view. This image forming device searches corresponding image information after scanning an original image, and if the corresponding image is found, outputs an image of the stored-image-information.
More specifically, Document 1 discloses a technique of image forming, which first carries out detection of characteristic quantity of scan-image-information, and searches for corresponding image information according to the characteristic quantity, and then carries out image forming based on the found image information. Document 2 discloses another matching technique, which carries out a vote for document files stored in a database to find a file with a descriptor identical to the descriptor denoting the extracted characteristic quantity of the input image, and the number of votes is counted. Further, Document 3 discloses a technique of carrying out search after the user selects document files by input from the keyword. Note that, in Document 3, the user is required to input the keyword (fraise) used for narrowing down the files, and therefore, depending on the selection of the keyword, the narrow-down of files may be carried out in an unwanted way.
The techniques of Documents 1 and 2 are not fully reliable in terms of accuracy when the characteristic quantity or the number of descriptors are small, and also the search will take a while when the characteristic quantity or the number of descriptors are large.
Further, the characteristic quantity or the descriptor only represent the feature of the original image, but are not the actual target images. Therefore, for example, if there are many image information items with similar characteristic quantities or descriptors, the foregoing search for desired image information may fail.
Further, if the search for the desired image is carried out by using the original image itself to find corresponding image information to the scan-image-information, the processing time significantly increases in proportion to the amount of original images.
Accordingly, those conventional methods are not useful to satisfy both improvement in searching accuracy of desired image information and quick search.
(Document 1)
Japanese Unexamined Patent Publication No. Tokukai 2001-16379(published on Jan. 19, 2001)
(Document 2)
Japanese Unexamined Patent Publication No. Tokukaihei 07-282088 (published on Oct. 27, 1995)
(Document 3)
Japanese Unexamined Patent Publication No. Tokukaihei 09-237282 (published on Sep. 9, 1997)